Progression to B acute lymphoblastic leukemia in 8p11 myeloproliferative syndrome with t(6;8)(q27;p12).

8p11 myeloproliferative syndrome is a rare hematological malignancy caused by the translocation of FGFR1. Patients present with a myeloproliferative neoplasm that frequently transforms into acute myeloid leukemia or T-lymphoblastic lymphoma/leukemia. Here, we report a molecular study of a patient with 8p11 myeloproliferative syndrome who developed acute B-lymphoblastic leukemia and then transformed to mixed-phenotype acute leukemia. A 67-year-old woman was diagnosed with a myeloproliferative neoplasm with t(6;8)(q27;p12) and was monitored for polycythemia vera. Four years later, she developed acute B-lymphoblastic leukemia with an additional chromosomal abnormality of - 7. Despite two induction regimens, she failed to achieve complete remission, and leukemia transformed into mixed-phenotype leukemia. Targeted sequencing of serial bone marrow samples identified the RUNX1 L144R mutation upon transformation to B-cell leukemia. After those two induction regimens, some RUNX1 mutation-positive leukemic cells obtained the JAK2 V617F mutation, which was associated with the emergence of myeloid markers, including myeloperoxidase.

MRI-Based Radiomics Differentiates Skull Base Chordoma and Chondrosarcoma: A Preliminary Study.

Chordoma and chondrosarcoma share common radiographic characteristics yet are distinct clinically. A radiomic machine learning model differentiating these tumors preoperatively would help plan surgery. MR images were acquired from 57 consecutive patients with chordoma (N = 32) or chondrosarcoma (N = 25) treated at the University of Tokyo Hospital between September 2012 and February 2020. Preoperative T1-weighted images with gadolinium enhancement (GdT1) and T2-weighted images were analyzed. Datasets from the first 47 cases were used for model creation, and those from the subsequent 10 cases were used for validation. Feature extraction was performed semi-automatically, and 2438 features were obtained per image sequence. Machine learning models with logistic regression and a support vector machine were created. The model with the highest accuracy incorporated seven features extracted from GdT1 in the logistic regression. The average area under the curve was 0.93 ± 0.06, and accuracy was 0.90 (9/10) in the validation dataset. The same validation dataset was assessed by 20 board-certified neurosurgeons. Diagnostic accuracy ranged from 0.50 to 0.80 (median 0.60, 95% confidence interval 0.60 ± 0.06%), which was inferior to that of the machine learning model (p = 0.03), although there are some limitations, such as the risk of overfitting and the lack of an extramural cohort for truly independent final validation. In summary, we created a novel MRI-based machine learning model to differentiate skull base chordoma and chondrosarcoma from multiparametric signatures.

Comparison of urethral sling surgery and non-ablative vaginal Erbium:YAG laser treatment in 327 patients with stress urinary incontinence: a case-matching analysis.

Stress urinary incontinence (SUI) occurs when abdominal pressure, such as from coughing or sneezing, causes urine leakage. We retrospectively compared tension-free vaginal tape (TVT) and non-ablative vaginal Erbium:YAG laser treatment (VEL) by propensity score (PS) analysis in women with SUI. No PS analysis studies have investigated urethral sling surgery using polypropylene TVT and VEL for SUI. Data from patients aged 35-50 years who were treated for SUI and registered at several institutions were selected. Patients with medical records covering 1 year for the 1-h pad test, who completed the International Consultation on Incontinence Questionnaire-Short Form (ICIQ-SF) and the Overactive Bladder Symptom Score (OABSS), were included. We analyzed 102, 113, and 112 patients in the TVT, VEL, and control groups, respectively. Compared with the control group, the TVT and VEL groups exhibited significant improvement in the 1-h pad test and ICIQ-SF. In the PS analysis, the TVT and VEL groups similarly improved in the 1-h pad test and ICIQ-SF. As for the OABSS, the VEL group showed significantly greater improvement than the TVT group. In the odds ratio analysis for the 1-h pad test, no differences in any of the parameters were observed between TVT and VEL. VEL may be considered an alternative to TVT for SUI treatment.

Challenging endovascular treatment for the patient with chronic limb threatening ischemia due to antiphospholipid syndrome.

Vascular surgery for chronic limb threatening ischemia (CLTI) in patients with antiphospholipid syndrome (APS) has a poor outcome with a high rate of postoperative thrombotic complications. However, there is little data regarding outcomes of endovascular treatment (EVT). This manuscript reports an interesting case of APS with CLTI in which timely EVT and continuous anticoagulant therapy not only prevented amputation but also thrombotic events over a long period of time. A 41-year-old man with CLTI in the left lower limb was seen and examined. Contrast-enhanced computed tomography revealed thrombotic occlusion in both iliac arteries and the left popliteal artery. Activated partial thromboplastin time was prolonged, and anti-ß2 glycoprotein I antibodies and lupus anticoagulants were present. After starting antithrombotic therapy with warfarin, EVT was performed using self-expandable stents in both iliac arterial lesions, and a marked reduction in limb ischemia was observed. Furthermore, the patient was free from thrombotic events for six years until a change in the anticoagulant led to stent thrombosis. This case suggests that EVT using stents along with continuous antithrombotic therapy with warfarin in patients with APS and CLTI could be very effective for immediate relief of ischemia and long-term reduction in thrombotic events. .

A case report of secondary neurolymphomatosis showing selective nerve infiltration and massive lumbar plexus enlargement.

BACKGROUND: Neurolymphomatosis (NL) is a rare manifestation of malignant lymphoma that shows selective infiltration to the peripheral nervous system primarily or secondarily. We report a patient with secondary NL caused by germinal center B-cell (GCB)-type diffuse large B-cell lymphoma (DLBCL) who showed selective infiltration of the lumbar plexus to the spinal cord and massive nerve enlargement resulting in severe pain. CASE PRESENTATION: A 72-year-old female exhibited asymmetric motor and sensory impairments and pain in the lower limbs that progressed for five months. Magnetic resonance imaging (MRI) showed an enlarged lumbar plexus, which continued to the cauda equina via the L3 and L4 spinal nerves. Her symptoms gradually worsened. Ten months after the onset of symptoms, the enlarged cauda equina filled the spinal canal space, and the spinal cord was swollen. A cauda equina biopsy was performed, and she was diagnosed with GCB-type DLBCL with CD10 positivity. The primary tumor was found in a mammary cyst. The autopsy study did not show apparent infiltration, except in the nervous system. CONCLUSIONS: Although there are many neurologic phenotypes of malignant lymphoma, the association between the cytological characteristics of lymphoma and the neurological phenotypes is still unclear. Several reports of CD10-positive secondary NL are available, whereas peripheral or central nervous tissue origin lymphoma cases are mostly negative for CD10. CD10 staining may be useful for distinguishing primary NL from secondary NL. NL often has a strong organotropism for peripheral nervous tissue, which makes early diagnosis challenging.

Development of Philadelphia chromosome-negative acute myeloid leukemia with IDH2 and NPM1 mutations in a patient with chronic myeloid leukemia who showed a major molecular response to tyrosine kinase inhibitor therapy.

Tyrosine kinase inhibitors (TKIs) are standard therapies for chronic myeloid leukemia (CML) that can eradicate Ph-positive leukemic cells. However, disease control is not achievable in a minority of cases, most commonly due to evolution of TKI-resistant clones. There have also been rare cases of emergence of Ph-negative clones with other cytogenetic abnormalities, and, less commonly, development of Ph-negative acute myeloid leukemia (AML), whose molecular pathogenesis is largely unknown. Here we report molecular features of a patient with Ph + CML who developed Ph-negative AML after showing a major molecular response to dasatinib. A 55-year-old man was diagnosed with CML. He achieved a complete cytogenetic response three months after dasatinib therapy but developed AML with normal karyotype 1 year later. After receiving induction and consolidation chemotherapy for AML, the patient achieved complete remission with no evidence of CML under maintenance with bosutinib. Targeted sequencing of serial bone marrow samples identified mutations in IDH2 and NPM1 in the Ph-negative AML cells, which had not been detected in CML cells. These results suggest that Ph-negative AML in this patient originated from a preleukemic population, which might have expanded during or after the successful elimination of CML clones with TKI therapy.

Stereotactic radiosurgery as a primary treatment for metastatic skull base alveolar soft part sarcoma: a case report.

Alveolar soft part sarcoma (ASPS) is a rare malignancy that typically arises in the trunk or extremities and preferentially metastasises to the brain. Radical resection is generally recommended for cranial metastatic ASPS, but stereotactic radiosurgery (SRS) is a recognised alternative for tumours in surgically challenging locations. Here, we present the case of a 22-year-old female, who underwent SRS and systemic therapy with pazopanib for a metastatic ASPS in the left temporal bone. The tumour was successfully controlled without further intervention over 23 months following SRS, which should be considered for metastatic ASPS when surgical resection is not appropriate.

Retrospective dose reconstruction of prostate stereotactic body radiotherapy using cone-beam CT and a log file during VMAT delivery with flattening-filter-free mode.

This study aimed to reconstruct the dose distribution of single fraction of stereotactic body radiotherapy for patients with prostate cancer using cone-beam computed tomography (CBCT) and a log file during volumetric-modulated arc therapy (VMAT) delivery with flattening-filter-free (FFF) mode. Twenty patients with clinically localized prostate cancer were treated with FFF-VMAT, and projection images for in-treatment CBCT (iCBCT) imaging were concomitantly acquired with a log file. A D95 dose of 36.25 Gy in five fractions was prescribed to each planning target volume (PTV) on each treatment planning CT (pCT). Deformed pCT (dCT) was obtained from the iCBCT using a hybrid deformable image registration algorithm. Dose distributions on the dCT were calculated using Pinnacle3 v9.10 by converting the log file data to Pinnacle3 data format using an in-house software. Dose warping was performed by referring to deformation vector fields calculated from pCT and dCT. Reconstructed dose distribution was compared with that of the original plan. Dose differences between the original and reconstructed dose distributions were within 3% at the isocenter and observed in PTV and organ-at-risk (OAR) regions. Differences in OAR regions were relatively larger than those in the PTV, presumably because OARs were more deformed than the PTV. Therefore, our method can be used successfully to reconstruct the dose distributions of one fraction using iCBCT and a log file during FFF-VMAT delivery.

[Effective treatment of POEMS syndrome accompanied by plasmacytoma with lenalidomide, dexamethasone, and local irradiation].

A 70-year-old woman experienced pain in both gastrocnemius muscles, numbness in the toes, and muscle weakness in both the legs that lasted for two months. After getting admitted to our hospital, the muscle weakness extended to both her arms, and nerve conduction studies revealed decreased nerve conduction velocity, which was more prominent in the elbow and the axilla than in the wrist. A magnetic resonance imaging revealed a tumor in the right femoral neck, which was histologically diagnosed as plasmacytoma. Laboratory findings revealed IgA lambda type M protein and an elevated VEGF level of 2,320 pg/ml; edema was present in both the legs. After a diagnosis of POEMS syndrome, lenalidomide and dexamethasone treatment were initiated simultaneously, along with irradiation. The treatment improved polyneuropathy, along with a decrease in the VEGF level. Increased vascular permeability due to elevated VEGF led to the development of neuropathy of POEMS syndrome, and treatment against proliferating monoclonal plasma cells is effective. In the present case, we believe that a prompt control of the plasmacytoma with novel therapeutic agents for myeloma with irradiation resulted in the improvement of the neurological symptoms.

Salvage stereotactic body radiotherapy for post-operative oligo-recurrence of non-small cell lung cancer: A single-institution analysis of 59 patients.

A standard treatment for patients with early-stage non-small cell lung cancer (NSCLC) who undergo surgery, and subsequently develop local failure or intrathoracic oligo-recurrence, has not yet been established. The present study aimed to assess the feasibility of stereotactic body radiotherapy (SBRT) for this subgroup of patients. Consequently, a retrospective analysis was conducted of patients with NSCLC recurrence who were treated with SBRT, and previously underwent curative surgical resection between October 2011 and October 2016. Post-SBRT survival [overall survival (OS); progression-free survival (PFS); and local control (LC)] and toxicity were analyzed. Prognostic factors for OS were identified using univariate and multivariate analysis. A total of 52 patients and 59 tumors were analyzed. The median follow-up time was 25 months (35 months for surviving patients), and median OS following salvage SBRT was 32 months. The 1- and 3-year OS rates were 84.4 and 67.8%, respectively. 1- and 3-year PFS rates were 80.8 and 58.7%, respectively. Only 4 patients (7.7%) developed local failure. Median LC was 71 months and 1- and 3-year LC rate were 97.9 and 94.9%, respectively. A total of 4 patients experienced grade 3 or higher adverse events (AEs) and two experienced grade 5 AEs (pneumonitis and hemoptysis). Central tumor location and the possibility of re-operation were independent prognostic factors for OS. The present study indicated that post-operative salvage SBRT is a promising therapeutic option for patients with NSCLC with locoregional or intrathoracic oligo-recurrence. We regard toxicity was also acceptable. However, further research is required on the appropriate selection of subjects, and stratification of the analysis by certain risk factors would increase the accuracy of the conclusions.

Publisher Correction: Prediction of malignant glioma grades using contrast-enhanced T1-weighted and T2-weighted magnetic resonance images based on a radiomic analysis.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Real-time markerless tumour tracking with patient-specific deep learning using a personalised data generation strategy: proof of concept by phantom study.

OBJECTIVE: For real-time markerless tumour tracking in stereotactic lung radiotherapy, we propose a different approach which uses patient-specific deep learning (DL) using a personalised data generation strategy, avoiding the need for collection of a large patient data set. We validated our strategy with digital phantom simulation and epoxy phantom studies. METHODS: We developed lung tumour tracking for radiotherapy using a convolutional neural network trained for each phantom's lesion by using multiple digitally reconstructed radiographs (DRRs) generated from each phantom's treatment planning four-dimensional CT. We trained tumour-bone differentiation using large numbers of training DRRs generated with various projection geometries to simulate tumour motion. We solved the problem of using DRRs for training and X-ray images for tracking using the training DRRs with random contrast transformation and random noise addition. RESULTS: We defined adequate tracking accuracy as the percentage frames satisfying <1 mm tracking error of the isocentre. In the simulation study, we achieved 100% tracking accuracy in 3 cm spherical and 1.5×2.25×3 cm ovoid masses. In the phantom study, we achieved 100 and 94.7% tracking accuracy in 3 cm and 2 cm spherical masses, respectively. This required 32.5 ms/frame (30.8 fps) real-time processing. CONCLUSIONS: We proved the potential feasibility of a real-time markerless tumour tracking framework for stereotactic lung radiotherapy based on patient-specific DL with personalised data generation with digital phantom and epoxy phantom studies. ADVANCES IN KNOWLEDGE: Using DL with personalised data generation is an efficient strategy for real-time lung tumour tracking.

Efficacy of a hydrogel spacer in three-dimensional conformal radiation therapy for prostate cancer.

OBJECTIVES: We aimed to compare the dose constraints fulfillment rate of the three-dimensional conformal radiotherapy treatment plan before and after a hydrogel spacer insertion. METHODS: The planning computed tomography scans of 39 patients who received stereotactic body radiotherapy for prostate cancer were used. All patients inserted a hydrogel spacer and underwent computed tomography scans before and after spacer insertion. The three-dimensional conformal radiotherapy plans according to NCCN classification, low-, intermediate- and high-risk, were made for each patient. Clinical target volume included prostate and seminal vesicle 2 cm for high risk, prostate and seminal vesicle 1 cm for intermediate risk and prostate only for low risk. Three-dimensional conformal radiotherapy including a seven-field conformal technique with 76 Gy in 38 fractions. Dose constraints for rectum and bladder were V70 Gy ≤ 15%, V65 Gy ≤ 30% and V40 Gy ≤ 60%. RESULTS: Among 39 patients, 35 (90%), 19 (49%) and 13 (33%) and 38 (97%), 38 (97%) and 34 (87%) patients before and after the spacer insertion fulfilled rectum dose constraints for low-, intermediate- and high-risk plans, respectively. A hydrogel spacer significantly reduced rectum dose and improved the rectum dose constraints fulfillment rate in intermediate (P < 0.01) and high (P < 0.01), but no difference was found in low-risk plan (P = 0.25). On multivariate analysis, spacer use was associated with the higher rectum dose constraints fulfillment rate. CONCLUSIONS: A hydrogel spacer reduced rectum dose and improved the dose constraints fulfillment rate in three-dimensional conformal radiotherapy plan. Although IMRT is the standard treatment, 3D-CRT using a hydrogel spacer may be a treatment option.

Long-term Outcomes of Gamma Knife Radiosurgery for Treating Vestibular Schwannoma With a Lower Prescription Dose of 12 Gy Compared With Higher Dose Treatment.

OBJECTIVE: Gamma knife radiosurgery (GKRS) is commonly used to treat vestibular schwannomas (VSs). The risk of complications from GKRS decreases at lower doses, but it is unknown if long-term tumor control is negatively affected by dose reduction. STUDY DESIGN: This was a retrospective case review and analysis of patient data. SETTING: Tertiary referral center. PATIENTS: Patients with VSs who underwent GKRS between 1990 and 2007 at the authors' institution. INTERVENTION(S): The subjects were divided into two cohorts based on the prescribed doses of radiation received: a 12 Gy cohort (96 patients) with a follow-up period of 124 months and a >12 Gy cohort (118 patients) with a follow-up period of 143 months. MAIN OUTCOME MEASURES: Tumor control rates at 10 to 15 years, frequency of facial and trigeminal nerve complications, and hearing function. RESULTS: The 10 to 15-year tumor control rates were 95% in the 12 Gy cohort and 88% in the > 12 Gy cohort, but the differences were not significant. Compared with the >12 Gy cohort, facial and trigeminal nerve deficits occurred significantly less frequently in the 12 Gy cohort, with the 10-year cumulative, permanent deficit-free rates being 2% and 0%, respectively. Multivariate analyses revealed that treatment doses exceeding 12 Gy were associated with a significantly higher risk for cranial nerve deficits. The percentage of subjects retaining pure-tone average ≤ 50 dB at the final follow-up did not significantly differ between the cohorts (12 Gy cohort, 30% and >12 Gy cohort, 33%; p = 0.823). CONCLUSIONS: Dose reduction to 12 Gy for GKRS to treat VSs decreased facial and trigeminal nerve complications without worsening tumor control rates.

Re-Evaluation of the Size Limitation in Single-Session Stereotactic Radiosurgery for Brain Arteriovenous Malformations: Detailed Analyses on the Outcomes with Focusing on Radiosurgical Doses.

BACKGROUND: Single-session stereotactic radiosurgery (SRS) for large arteriovenous malformations (AVMs) ≥10 mL remains controversial, which is considered as the current size limitation. OBJECTIVE: To reconsider the size limitation of SRS for AVMs by profoundly analyzing dose-volume relationship. METHODS: Data on 610 consecutive patients with AVM treated with SRS using regular (18-22 Gy) or low (<18 Gy) prescription doses were retrospectively analyzed. AVMs were classified into 4 groups: small (<5 mL), medium (≥5 and <10 mL), medium-large (≥10 and <15 mL), and large (≥15 mL). The maximum volumes were 22.5 mL (regular-dose group) and 23.5 mL (low-dose group). RESULTS: When treated with regular doses, the cumulative 6-yr obliteration rates for each of the 4 AVM groups were 86%, 80%, 87%, and 79%, respectively; the cumulative 10-yr significant neurological event (SNE) rates were 2.6%, 3.9%, 6.8%, and 5.3%, respectively. Regarding large AVMs, regular-dose SRS resulted in marginally better obliteration rate (6-yr cumulative rate, 79% vs 48%, P = .111) and significantly lower SNE (5-yr cumulative rate, 5% vs 31%, P = .038) and post-SRS hemorrhage rate (8-yr cumulative rate, 0% vs 54%, P = .002) compared to low-dose SRS. Multivariate analyses revealed that regular-dose SRS significantly contributed to increase in the obliteration rate and decrease in SNEs and hemorrhage. CONCLUSION: The outcomes for large AVMs were generally favorable when treated with ablative doses. Single-session SRS could be acceptable for AVMs up to ≈20 mL if treated with ablative doses.

Acceptable fetal dose using flattening filter-free volumetric arc therapy (FFF VMAT) in postoperative chemoradiotherapy of tongue cancer during pregnancy.

Optimizing irradiation protocols for pregnant women is challenging, because there are few cases and a dearth of fetal dosimetry data. We cared for a 36-year-old pregnant woman with tongue cancer. Prior to treatment, we compared three intensity-modulated radiation therapy (IMRT) techniques, including helical tomotherapy, volumetric arc therapy (VMAT), and flattening-filter free VMAT (FFF-VMAT) using treatment planning software. FFF-VMAT achieved the minimum fetal exposure and was selected as the optimal modality. We prescribed 66 Gy to the involved nodes, 60 Gy to the tumor bed and ipsilateral neck, and 54 Gy to the contralateral neck over 33 fractions. To confirm the out-of-field exposure per fraction, surface doses and the rectal dose were measured during FFF-VMAT delivery. Postoperative chemoradiotherapy was delivered using IMRT and a cisplatin regimen. Without any shielding, the total fetal dose was 0.03 Gy, within the limits established by the ICRP. A healthy girl was born vaginally at 37 weeks' gestation.

Prediction of malignant glioma grades using contrast-enhanced T1-weighted and T2-weighted magnetic resonance images based on a radiomic analysis.

We conducted a feasibility study to predict malignant glioma grades via radiomic analysis using contrast-enhanced T1-weighted magnetic resonance images (CE-T1WIs) and T2-weighted magnetic resonance images (T2WIs). We proposed a framework and applied it to CE-T1WIs and T2WIs (with tumor region data) acquired preoperatively from 157 patients with malignant glioma (grade III: 55, grade IV: 102) as the primary dataset and 67 patients with malignant glioma (grade III: 22, grade IV: 45) as the validation dataset. Radiomic features such as size/shape, intensity, histogram, and texture features were extracted from the tumor regions on the CE-T1WIs and T2WIs. The Wilcoxon-Mann-Whitney (WMW) test and least absolute shrinkage and selection operator logistic regression (LASSO-LR) were employed to select the radiomic features. Various machine learning (ML) algorithms were used to construct prediction models for the malignant glioma grades using the selected radiomic features. Leave-one-out cross-validation (LOOCV) was implemented to evaluate the performance of the prediction models in the primary dataset. The selected radiomic features for all folds in the LOOCV of the primary dataset were used to perform an independent validation. As evaluation indices, accuracies, sensitivities, specificities, and values for the area under receiver operating characteristic curve (or simply the area under the curve (AUC)) for all prediction models were calculated. The mean AUC value for all prediction models constructed by the ML algorithms in the LOOCV of the primary dataset was 0.902 ± 0.024 (95% CI (confidence interval), 0.873-0.932). In the independent validation, the mean AUC value for all prediction models was 0.747 ± 0.034 (95% CI, 0.705-0.790). The results of this study suggest that the malignant glioma grades could be sufficiently and easily predicted by preparing the CE-T1WIs, T2WIs, and tumor delineations for each patient. Our proposed framework may be an effective tool for preoperatively grading malignant gliomas.

Long-Term Outcomes of Stereotactic Radiosurgery for Vestibular Schwannoma Associated with Neurofibromatosis Type 2 in Comparison to Sporadic Schwannoma.

The efficacy of radiosurgery for neurofibromatosis type 2 (NF2)-associated vestibular schwannoma (VS) remains debatable. We retrospectively analyzed radiosurgical outcomes for NF2-associated VS compared to sporadic VS using our database of 422 consecutive VS patients. Twenty-five patients with 30 NF2-associated VSs with a mean follow-up of 121 months were identified. NF2-associated VSs exhibited excellent tumor control (10-year cumulative rate, 92% vs. 92% in sporadic VSs; p = 0.945) and worse overall survival (73% vs. 97%; p = 0.005), mainly due to tumor progression other than the treated VSs. The presence of NF2 was not associated with failed tumor control via multivariate Cox proportional hazard analyses. No difference in radiation-induced adverse events (RAEs) was confirmed between cohorts, and prescription dose (hazard ratio 8.30, 95% confidence interval 3.19-21.62, p < 0.001) was confirmed as a risk for cranial nerve injuries via multivariate analysis. Further analysis after propensity score matching using age, volume, and sex as covariates showed that NF2-associated VSs exhibited excellent local control (100% vs. 93%; p = 0.240) and worse overall survival (67% vs. 100%; p = 0.002) with no significant difference in RAEs. Excellent long-term tumor control and minimal invasiveness may make radiosurgery a favorable therapeutic option for NF2 patients with small to medium VS, preferably with non-functional hearing or deafness in combination with postoperative tumor growth or progressive non-operated tumors, or with functional hearing by patients' wish.

Radiomics Analysis for Glioma Malignancy Evaluation Using Diffusion Kurtosis and Tensor Imaging.

PURPOSE: A noninvasive diagnostic method to predict the degree of malignancy accurately would be of great help in glioma management. This study aimed to create a highly accurate machine learning model to perform glioma grading. METHODS AND MATERIALS: Preoperative magnetic resonance imaging acquired for cases of glioma operated on at our institution from October 2014 through January 2018 were obtained retrospectively. Six types of magnetic resonance imaging sequences (T2-weighted image, diffusion-weighted image, apparent diffusion coefficient [ADC], fractional anisotropy, and mean kurtosis [MK]) were chosen for analysis; 476 features were extracted semiautomatically for each sequence (2856 features in total). Recursive feature elimination was used to select significant features for a machine learning model that distinguishes glioblastoma from lower-grade glioma (grades 2 and 3). RESULTS: Fifty-five data sets from 54 cases were obtained (14 grade 2 gliomas, 12 grade 3 gliomas, and 29 glioblastomas), of which 44 and 11 data sets were used for machine learning and independent testing, respectively. We detected 504 features with significant differences (false discovery rate <0.05) between glioblastoma and lower-grade glioma. The most accurate machine learning model was created using 6 features extracted from the ADC and MK images. In the logistic regression, the area under the curve was 0.90 ± 0.05, and the accuracy of the test data set was 0.91 (10 out of 11); using a support vector machine, they were 0.93 ± 0.03 and 0.91 (10 out of 11), respectively (kernel, radial basis function; c = 1.0). CONCLUSIONS: Our machine learning model accurately predicted glioma tumor grade. The ADC and MK sequences produced particularly useful features.